Question

Your team is running Vertex AI hyperparameter tuning on a deep learning model that takes about 6 hours per trial. You have a fixed budget of 40 total trials and want the best possible objective value. You are using Bayesian optimization as the search algorithm. Your manager wants results delivered as quickly as possible, but also wants the search to be as effective as possible. How should you configure the maxParallelTrials setting to balance wall-clock time and search quality?

Accepted Answer

Set maxParallelTrials to a moderate value (e.g., 4-8) so Bayesian optimization can still learn from completed trials while achieving reasonable parallelism.

Answer

Set maxParallelTrials equal to maxTrialCount (40) so all trials run simultaneously and finish in the shortest wall-clock time.

Answer

Set maxParallelTrials to 1 so every trial fully benefits from all prior results, maximizing search quality regardless of time.

Answer

Switch the algorithm to grid search and set maxParallelTrials to 40, since grid search parallelizes without any quality penalty.